Re-Learning to Be Different: Neural Differentiation Supports Post-stroke Language Recovery
Jeremy Purcell1,2, Robert Wiley1, Brenda Rapp1; 1Department of Cognitive Science, Johns Hopkins University, USA, 2Maryland Neuroimaging Center, University of Maryland, USA
Little is known about the changes in neural representations that support post-stroke recovery. Recent work suggests that the local differentiation of neural responses reflects representational integrity and learning, with differentiation increasing with expertise and learning (e.g., Jiang et al., 2013). We apply a novel technique – Local-Heterogeneity Regression (Local-Hreg) - to examine neural representations before and after behavioral treatment in 20 individuals with acquired dysgraphia due to stroke. For treatment, individuals were trained to spell an individualized word list for approximately 3 months. FMRI with a spelling task was carried out pre- and post-training. A whole-brain, Local-Hreg searchlight and traditional GLM analysis were performed. Local-Hreg measures local neural differentiation by quantifying the relative dissimilarity in the BOLD response across adjacent voxels within a search light. Overall, we found converging evidence that high neural differentiation in the left ventral occipitotemporal cortex (vOTC) was related to better performance and predicted future improvements due to treatment. We also found that there were selective increases in neural differentiation within the left vOTC and that the amount of increase was related to the magnitude of behavioral improvements. Finally, we did not observe any changes in mean BOLD response within the left vOTC, nor did we observe any relationship with behavior. This work provides a novel approach for quantifying neural re-differentiation of local representations, and reveals that this measure can be used to index performance, predict response to treatment and quantify neural changes due to re-learning in post-stroke recovery.
Topic Area: LANGUAGE: Lexicon